Local linear density estimators achieve automatic boundary corrections and enjoy some typical optimal properties. Proper choice of the smoothing parameters is crucial for their performance. A data-based bandwidth selector is developed in the spirit of plug-in rules. Consistency and asymptotic normality of the selected bandwidth are demonstrated. The bandwidth is very efficient regardless of whether there are non-smooth boundaries in the support of the density or not.
"A bandwidth selector for local linear density estimators." Ann. Statist. 25 (3) 1001 - 1013, June 1997. https://doi.org/10.1214/aos/1069362735